JDINAC: joint density-based non-parametric differential interaction network analysis and classification using high-dimensional sparse omics data

摘要: Motivation: A complex disease is usually driven by a number of genes interwoven into networks, rather than single gene product. Network comparison or differential network analysis has become an important means revealing the underlying mechanism pathogenesis and identifying clinical biomarkers for classification. Most studies, however, are limited to correlations that mainly capture linear relationship among genes, rely on assumption parametric probability distribution measurements. They restrictive in real application. Results: We propose new Joint density based non-parametric Differential Interaction Analysis Classification (JDINAC) method identify interaction patterns activation between two groups. At same time, JDINAC uses build classification model. The novelty lies its potential non-linear relations molecular interactions using high-dimensional sparse data as well adjust confounding factors, without need Simulation studies demonstrate provides more accurate estimation lower error achieved other state-of-the-art methods. apply Breast Invasive Carcinoma dataset, which includes 114 patients who have both tumor matched normal samples. hub identified were consistent with existing experimental studies. Furthermore, discriminated sample high accuracy virtue biomarkers. general framework feature selection omics data.